Stefaan Verhulst

Blog by Fernando Monge: “In 2020, Wellington’s city council decided to map 16 kilometers of the city’s streets using GPR and Lidar technologies. They were astounded by what they found. Just in that relatively shorter mileage they identified 100 anomalies – including a collapsed water main. This pilot and an additional study gave them the justification they needed to secure the resources – 4 million dollars from the central government – to launch the pilot for an underground asset map.1

Identifying a need and getting funding to tackle it is one thing. Coordinating across the infrastructure sector is another. Aware of the importance of collaboration for the project’s success, the city council established a technical reference group with representatives from the city, the water entity, electricity and gas distributors, telecommunication providers, contractors and other experts.

Another core piece of the governance framework was finding a neutral entity that would manage the register. This is key in any data sharing scheme. Companies and public entities sharing the data want to trust that whomever is receiving their data will not share it with anyone who shouldn’t have access to it. They also need to be confident of technical security aspects. In the case of the UAR, this intermediary role was played by Digital Built Aotearoa, a neutral not‑for‑profit entity that operates the platform and runs a federated data sharing scheme, ensuring that no one single entity controls all the data.

Through this collaboration process – one of the most interesting aspects of the UAR’s experience – partners agreed on the type of information, standards and sharing mechanisms of the UAR. For example, assets’ attributes such as type (water main, sewer, power cable…), ownership, dimension, material, installation data and positional accuracy need to be included in the register. Metadata also describes how each dataset was collected, its refresh rate and usage constraints. See an example of some of the data collected per asset:

The quality of the data is determined based on whether it was derived from existing records or precisely surveyed. Asset owners share it with the platform through APIs or file uploads. Field crews of asset owners or construction companies can also share data directly via a mobile app when they find something while digging…(More)”.

Book by John P. Wihbey: “The large, corporate global platforms networking the world’s publics now host most of the world’s information and communication. Much has been written about social media platforms, and many have argued for platform accountability, responsibility, and transparency. But relatively few works have tried to place platform dynamics and challenges in the context of history, especially with an eye toward sensibly regulating these communications technologies. In Governing Babel, John Wihbey considers the ongoing, high-stakes debate over social media platforms and free speech, and how these companies ought to manage their tremendous power.

Wihbey takes readers on a journey into the high-pressure and controversial world of social media content moderation, looking at issues through relevant cultural, legal, historical, and global lenses. The book addresses a vast challenge—how to create new rules to deal with the ills of our communications and media systems—but the central argument it develops is relatively simple. The idea is that those who create and manage systems for communications hosting user-generated content have both a responsibility to look after their platforms and a duty to respond to problems. They must, in effect, adopt a central response principle that allows their platforms to take reasonable action when potential harms present themselves. And finally, they should be judged, and subject to sanction, according to the good faith and persistence of their efforts…(More)”.

Article by Eli Tan: “The devices are part of an industry known as precision farming, a data-driven approach for optimizing production that is booming with the addition of A.I. and other technologies. Last year, the livestock-monitoring industry alone was valued at more than $5 billion, according to Grand View Research, a market research firm.

Farmers have long used technology to collect and analyze data, with the origins of precision farming dating to the 1990s. In the early 2000s, satellite imagery changed the way farmers determined crop schedules, as did drones and eventually sensors in the fields. Nowadays, if you drive by farms in places like California’s Central Valley, you may not see any humans at all.It is not just dairy farms seeing a change. Elsewhere in the Central Valley, which produces about half of America’s fruits and vegetables, autonomous tractors that use the same sensors as robot taxis and apple-picking tools with A.I. cameras have become popular.

The most common precision farming technology, like GPS maps that track crop yields and auto-steering tractors, are used by 70 percent of large farms today, up from less than 10 percent in the early 2000s, according to Department of Agriculture data.

“The tech is getting faster every year,” said Deepak Joshi, a professor of precision agriculture at Kansas State University. “It used to be we’d have new technologies every few years, and now it’s every six months.”The new products are helping farmers reduce costs as tariffs and inflation raise the prices of farm equipment and feed. They also allow farmers to do more work with fewer people as the Trump administration cracks down on illegal immigration.

Precision farming has boomed as the costs of cameras and sensors have plunged and A.I. models that analyze data have improved, said Charlie Wu, the founder of Orchard Robotics, a farming robotics start-up in San Francisco. Much like the rest of the A.I. industry, agricultural technology is now anchored by chips made by Nvidia, he said…(More)”.

Book by Sigge Winther Nielsen: “The wicked problems of our time – climate change, migration, inequality, productivity, and mental health – remain stubbornly unresolved in Western democracies. It’s time for a new way to puzzle.

This book takes the reader on a journey from London to Berlin, and from Rome to Washington DC and Copenhagen, to discover why modern politics keeps breaking down. We’re tackling twenty-first century problems with twentiethcentury ideas and nineteenth-century institutions.

In search of answers, the author gate-crashes a Conservative party at an English estate, visits a stressed-out German minister at his Mallorcan holiday home, and listens in on power gossip in Washington DC’s restaurant scene. The Puzzle State offers one of the most engaging and thoughtful analyses of Western democracies in recent years. Based on over 100 interviews with top political players, surveys conducted across five countries, and stories of high-profile policy failures, it uncovers a missing link in reform efforts: the absence of ongoing feedback loops between decision-makers and frontline practitioners in schools, hospitals, and companies.

But there is a way forward. The author introduces the concept of the Puzzle State, which uses collective intelligence and AI to (re)connect politicians with the people implementing reforms on the ground. The Puzzle State tackles highprofile wicked problems and enables policies to adapt as they meet messy realities. No one holds all the pieces in the Puzzle State. The feedback loop must cut across all sectors – from civil society to corporations – just like solving a complex puzzle requires commitment, cooperation, and creativity…(More)”.

Article by Jenifer Whiston: “At Free Law Project, we believe the law belongs to everyone. But for too long, the information needed to understand and use the law—especially in civil rights litigation—has been locked behind paywalls, scattered across jurisdictions, or buried in technical complexity.

That’s why we teamed up with the Civil Rights Litigation Clearinghouse on an exploratory grant from Arnold Ventures: to see how artificial intelligence could help unlock these barriers, making court documents more accessible and legal research more open and accurate.

What We Learned

This six-month project gave us the opportunity to experiment boldly. Together, we researched and tested AI based approaches and tools to:

• Classify legal documents into categories like motions, orders, or opinions.
• Summarize filings and entire cases, turning dense legal text into plain-English explanations.
• Generate structured metadata to make cases easier to track and analyze.
• Trace the path of a legal case as it moves through the court system
• Enable semantic search, allowing users to type questions in everyday language and find relevant cases.
• Build the foundation of an open-source citator, so researchers and advocates can see when cases are overturned, affirmed, or otherwise impacted by later decisions.

These prototypes showed us not only what’s possible, but what’s practical. By testing real AI models on real data, we proved that tools like these can responsibly scale to improve the infrastructure of justice…(More)”.

Report by Rainer Kattel et al: “The four-stage assessment approach for assessing city government dynamic capabilities reflects the complexity of city governments. It blends rigour with usefulness, striving to be both robust and usable, adding value to city governments and those that support their efforts. This report summarises how to assess and compare dynamic capabilities across city governments, provides an assessment of dynamic capabilities in a selected sample set of cities, and sets out future work that will explore how to create an index that can be scaled to thousands of cities and used to effectively build capabilities, positively transforming cities and creating better lives for residents…(More)”.

Blog by the Stanford Peace Innovation Lab: “…In the years since, the concept of peace technology has expanded globally. Accelerators—structures that concentrate resources, mentorship, and networks around innovation—have played a central role in this evolution. Governments, universities, nonprofits, and private companies now host peace tech accelerators, each bringing their own framing of what “peace tech” means. Looking across these initiatives, we see the emergence of a diverse but connected ecosystem that, as of 2025, continues to grow with new conferences, alliances, and prizes…(More)”.

Paper by Lukas Linsi et al: “International organisations (IOs) collect and disseminate a wide array of statistics. In many cases, the phenomena these statistics seek to quantify defy precise measurement. Hard-to-measure statistics frequently represent ‘guesstimates’ rather than solid measures. Nonetheless, they are often presented as if they were perfectly reliable, precise estimates. What drives IOs to disseminate guesstimates and why are they frequently presented with seemingly excessive precision? To answer these questions, we adopt an ecological framework where IOs must pay attention to both internal and external audiences. The framework informs three mechanisms: Mock precise guesstimates are fuelled by how organisations seek to attract attention to their work, signal scientific competence, and consolidate their professional standing. Empirically, we evaluate these mechanisms in the context of statistics on (illicit) trade, employing a mixed methods approach that integrates elite interviews, expert surveys and a survey experiment. Our findings show how organisational and professional incentives lead to the use of mock precision. While the field we study is IOs, the underlying dynamics are of broader applicability. Not least, they raise questions about mock precision as a (un)scientific practice also commonly used by academic researchers in international political economy (IPE), economics and the wider social sciences…(More)”.

Report by BVBA Research: “…presents a novel method to track technological progress in real time using data from ArXiv, the world’s most popular preprint repository in computer science, physics and mathematics.

Key points

• Understanding the process of science creation is key to monitoring the rapid digital transformation. Using natural language processing techniques (NLP), at BBVA Research we have developed a new set of technological indicators using data from ArXiv, the world’s most popular open-access preprint repository in physics, mathematics and computer science.
• Our monthly indicators track ArXiv preprint publications in computer science, physics, and mathematics to deliver a granular and real-time signal of research activity, thus of technological progress and digital transformation. These metrics highly correlate with traditional proxies of research activity such as patents.
• Computer science preprints have exploded from under 10% of all ArXiv submissions in 2010 to nearly 50 % in 2025, driven predominantly by AI research now representing about 70% of all computer science output. Conversely, research on physics and mathematics have grown in parallel at a steady rate.
• There was a turning point in the computer science paradigm around 2014, with the research focus drastically moving from classic computing theory (e.g. information systems, data structures and numerical analysis) to AI-related disciplines (machine learning, computer vision, computational linguistics and others). Research on cryptography and security was the only subfield essentially unaffected by the AI boom.
• Scientific innovation plays a crucial role in shaping productivity trends, labor market dynamics, capital flows, investment strategies, and regulatory policies. These indicators provide valuable insights into our rapidly evolving world, enabling early detection of technological shocks and informing strategic investment decisions…(More)”.

Essay by Ed Bradon: “Systems thinking promises to give us a toolkit to design complex systems that work from the ground up. It fails because it ignores an important fact: systems fight back…

The systems that enable modern life share a common origin. The water supply, the internet, the international supply chains bringing us cheap goods: each began life as a simple, working system. The first electric grid was no more than a handful of electric lamps hooked up to a water wheel in Godalming, England, in 1881. It then took successive decades of tinkering and iteration by thousands of very smart people to scale these systems to the advanced state we enjoy today. At no point did a single genius map out the final, finished product.

But this lineage of (mostly) working systems is easily forgotten. Instead, we prefer a more flattering story: that complex systems are deliberate creations, the product of careful analysis. And, relatedly, that by performing this analysis – now known as ‘systems thinking’ in the halls of government – we can bring unruly ones to heel. It is an optimistic perspective, casting us as the masters of our systems and our destiny…(More)”.